def build(self, moleculeData=None, name=None, unit='angstrom', **kwargs):
if moleculeData is not None:
if type(moleculeData) is list:
moleculeData = np.array(moleculeData)
if len(moleculeData.shape) == 1:
moleculeData = np.atleast_2d(moleculeData)
self.N = moleculeData.shape[0]
self.Z = moleculeData[:, 0]
self.R = moleculeData[:, 1:]
self.type_list = [qtk.Z2n(z) for z in self.Z]
self.string = ['' for i in range(self.N)]
if name is None:
self.name = self.stoichiometry()
else:
self.name = name
else:
assert 'Z' in kwargs
assert 'R' in kwargs
Z = np.array(kwargs['Z'])
R = np.array(kwargs['R'])
self.N = len(Z)
self.Z = Z
self.R = R
self.type_list = [qtk.Z2n(z) for z in self.Z]
self.string = ['' for i in range(self.N)]
self.name = self.stoichiometry()
if unit is 'bohr':
self.R = self.R * 0.52917721092
return self
def stoichiometry(self, output='string', **kwargs):
elements = collections.Counter(sorted(self.Z))
data = zip(elements.keys(), elements.values())
data.sort(key=lambda tup: tup[0])
if output == 'dictionary' or output == 'count':
out = {}
for element in data:
out[qtk.Z2n(element[0])] = element[1]
elif output == 'string':
out = ''
for element in data:
out = out + qtk.Z2n(element[0]) + str(element[1])
return out
def _mutate(self, mutation):
for m in xrange(len(mutation)):
for i in xrange(len(mutation[m])):
index = self.mutation_list[m][i]
#target = self.mutation_target[m][mutation[m][i]]
target = mutation[m][i]
self.new_structure.Z[index] = target
self.new_structure.type_list[index] = qtk.Z2n(target)
def mutateElement(self, oldZ, newZ):
if oldZ != newZ:
if type(oldZ) is str:
oldZ = qtk.n2Z(oldZ)
if type(newZ) is str:
newZ = qtk.n2Z(newZ)
newElem = qtk.Z2n(newZ)
indices = []
for i in range(self.N):
if abs(self.Z[i] - oldZ) < 0.1:
indices.append(i)
self.setAtoms(indices, element=newElem)
def setAtoms(self, index, **kwargs):
if type(index) is int:
index = [index]
if 'element' in kwargs or 'Z' in kwargs:
for i in index:
if 'element' in kwargs:
if type(kwargs['element']) is str:
Z = qtk.n2Z(kwargs['element'])
Zn = kwargs['element']
elif type(kwargs['element']) is int\
or type(kwargs['element']) is float:
Z = kwargs['element']
Zn = qtk.Z2n(Z)
elif 'Z' in kwargs:
Z = kwargs['Z']
Zn = qtk.Z2n(Z)
self.Z[i] = Z
self.type_list[i] = Zn
if 'string' in kwargs:
minZ = min(min(self.Z)-1, 0)
for i in index:
self.string[i] = kwargs['string']
self.Z[i] = minZ
def write(self, name=None):
today = datetime.date.today()
out = InpContent(name)
out.write('Goedecker pseudopotential for %s\n' % \
qtk.Z2n(self.param['Z']))
out.write(' %2d %2d %s ! Z, ZV, date(ddmmyy)\n' %\
(self.param['Z'], self.param['ZV'], today.strftime('%d%m%y')))
out.write(' %d %s 2 0 2002 0 ! xc:%s\n' %\
(type_dict[self.setting['type']],
xc_dict[self.param['xc']],
self.param['xc']))
out.write(' %12.8f %3d' % (self.param['r_loc'], self.param['Cn']))
for i in range(self.param['Cn']):
out.write(' % 12.8f' % self.param['Ci'][i])
out.write(' ! rloc, #C, C[i]')
out.write('\n %d\n' % self.param['l_max'])
for i in range(len(self.param['r_nl'])):
r_nl = self.param['r_nl'][i]
h_ij = self.param['h_ij'][i]
out.write(' %12.8f %3d' % (r_nl, len(h_ij)))
if len(h_ij) > 0:
for j in range(len(h_ij)):
if j > 0:
out.write('%19s' % '')
for k in range(j, len(h_ij)):
out.write(' % 12.8f' % h_ij[j][k])
if j == 0:
out.write(' ! h_ij\n')
else:
out.write('\n')
else:
out.write(' % 12.8f\n' % 0)
if self.param['r_nl'] > 0:
for j in range(len(h_ij)):
out.write('%19s' % '')
for k in range(j, len(h_ij)):
out.write(' % 12.8f' % 0.0)
if j == 0:
out.write(' ! abinit k_ij\n')
else:
out.write('\n')
if self.setting['type'] == 'nlcc':
out.write(' %12.8f %12.8f' %\
(self.param['rcore'], self.param['qcore']))
out.close()
def ESP(self, coord=None, **kwargs):
"""
method for electron density
Note: CUBE file is assumed to be orthorohmbic
"""
data = self.data
grid = self.grid
if 'molecule' not in kwargs:
mol = self.molecule
else:
try:
mol = copy.deepcopy(qtk.toMolecule(kwargs['molecule']))
except:
qtk.exit("error when loading molecule:%s" % \
str(kwargs['molecule']))
N = mol.N
Q = self.integrate()
Z_sum = sum(mol.Z)
ne_diff = abs(Q - Z_sum)
ve_diff = abs(Q - mol.getValenceElectrons())
if min(ne_diff, ve_diff) > 1E-2:
qtk.warning("charge not conserved... ESP is wrong!")
qtk.warning("charge integrate to %.2f, " % Q + \
"while nuclear charge adds to %.2f" % Z_sum)
if ve_diff < ne_diff:
Z = [qtk.n2ve(qtk.Z2n(z)) for z in mol.Z]
Z = np.array(Z).reshape(N, 1)
else:
Z = mol.Z.reshape(N, 1)
structure = np.hstack([Z, mol.R * 1.889725989])
if coord is not None:
x, y, z = np.array(coord) * 1.889725989
V = ESP_c.esp_point(grid, structure, data, x, y, z)
return V
else:
qtk.warning("known issue: unidentifed memory leak")
out = copy.deepcopy(self)
out.molecule = mol
out.data = np.nan_to_num(ESP_cube_c.esp_cube(
grid, structure, data))
return out
def addAtoms(self, element, coord):
if type(element) is not list:
element = [element]
coord = [coord]
if type(element[0]) is not str:
element = [qtk.Z2n(int(i)) for i in element]
Z = list(self.Z)
for i in range(len(element)):
#e = getattr(pt, element[i].title())
e = qtk.element[element[i]]
r = coord[i]
if self.N == 0:
self.R = np.array(r)
else:
self.R = np.vstack([self.R, np.array(r)])
self.N = self.N + 1
self.type_list.append(e.symbol)
Z.append(e.number)
self.string.append('')
self.Z = np.array(Z)
if self.N == 1:
self.R = np.array([self.R])
def getMO(self, fchk):
self.program = 'gaussian'
fchkfile = open(fchk)
fchk = fchkfile.readlines()
self.basis_name = filter(None, fchk[1].split(' '))[2]
def basisList(L):
N_dict = {4: 'g', 5: 'h', 6: 'k'}
orbital = ['x', 'y', 'z']
base = [2 for _ in range(L)]
out = []
for n in range(L + 1):
b = copy.deepcopy(base)
for i in range(n):
b[i] = 0
orb = N_dict[L] + ''.join([orbital[j] for j in b])
out.append(orb)
for i in range(n, L):
b[i] = 1
orb = N_dict[L] + ''.join([orbital[j] for j in b])
out.append(orb)
return out
basis_list = [
['s'],
['px', 'py', 'pz'],
['dxx', 'dyy', 'dzz', 'dxy', 'dxz', 'dyz'],
[
'fxxx', 'fyyy', 'fzzz', 'fxyy', 'fxxy', 'fxxz', 'fxzz', 'fyzz',
'fyyz', 'fxyz'
],
basisList(4),
basisList(5),
]
def readFchk(flag, type=float):
flagStr = filter(lambda x: flag in x, fchk)[0]
n_entry = int(filter(None, flagStr.split(' '))[-1])
ind = fchk.index(flagStr) + 1
if type is float:
factor = 5
elif type is int:
factor = 6
n_lines = n_entry / factor + 1
if not n_entry % factor: n_lines = n_lines - 1
data = []
for i in range(ind, ind + n_lines):
dList = list(
np.array(filter(None, fchk[i].split(' '))).astype(type))
data.extend(dList)
return data, n_entry
self.Z, self.N = readFchk('Nuclear charges')
self.type_list = [qtk.Z2n(z) for z in self.Z]
_types, _n_shell = readFchk('Shell types', int)
_exp, _nfn = readFchk('Primitive exponents')
_cef = readFchk('Contraction coefficients')[0]
_coord, _test = readFchk('Coordinates of each shell')
_ng = readFchk('Number of primitives per shell', int)[0]
_coord = np.array(_coord).reshape([_n_shell, 3])
self.mo_eigenvalues, self.n_mo = \
readFchk('Alpha Orbital Energies')
_mo, _dim = readFchk('Alpha MO coefficients')
try:
self.mo_eigenvalues_beta, _ = \
readFchk('Beta Orbital Energies')
_mob, _dimb = readFchk('Beta MO coefficients')
except:
_mob, _dimb = None, None
self.mo_eigenvalues_beta = None
self.n_ao = _dim / self.n_mo
self.n_basis = self.n_ao
self.mo_vectors = np.array(_mo).reshape([self.n_mo, self.n_ao])
if self.mo_eigenvalues_beta is not None:
self.mo_vectors_beta = np.array(_mob).reshape(
[self.n_mo, self.n_ao])
_map = readFchk('Shell to atom map', int)[0]
_map_coord = list(np.diff(_map))
_map_coord.insert(0, 1)
_R_ind = [i for i in range(_n_shell) if _map_coord[i] > 0]
self.R_bohr = np.array([_coord[i] for i in _R_ind])
self.R = self.R_bohr / 1.88972613
_neStr = filter(lambda x: 'Number of electrons' in x, fchk)[0]
_ne = float(filter(None, _neStr.split(' '))[-1])
self.occupation = []
warned = False
for i in range(self.n_mo):
if i < _ne / 2 - 1:
self.occupation.append(2.0)
elif i >= _ne / 2 - 1 and i < _ne / 2:
if abs(_ne % 2) < 1E-5:
self.occupation.append(2.0)
else:
self.occupation.append(_ne % 2)
else:
self.occupation.append(0.0)
tmp = 0
shift = 0
self.basis = []
for i in range(_n_shell):
_ind = _map[i] - 1
bfn_base = {}
bfn_base['index'] = i
bfn_base['center'] = _coord[i]
bfn_base['atom'] = self.type_list[_ind]
exp = []
cef = []
for g in range(_ng[i]):
exp.append(_exp[i + shift + g])
cef.append(_cef[i + shift + g])
shift = shift + _ng[i] - 1
if _types[i] == 0:
l_max = 1
blist = ['s']
else:
if _types[i] > 0:
l_max = sum(range(_types[i] + 2))
blist = basis_list[_types[i]]
else:
l_max = 2 * abs(_types[i]) + 1
blist = ['D0' for _ in range(l_max)]
if not warned:
qtk.warning("sp shell or spherical basis are used, " +\
"some functions will not work")
warned = True
for l in range(l_max):
bfn = copy.deepcopy(bfn_base)
bfn['exponents'] = copy.deepcopy(exp)
bfn['coefficients'] = copy.deepcopy(cef)
bfn['type'] = blist[l]
tmp = tmp + 1
self.basis.append(bfn)
self.basisFormat()
def getChild():
child = {}
for key in parent1.iterkeys():
if key == 'mutation':
# implementation for element_count constraints
if self.element_count:
constraint_list = []
constraint_keys = []
# tool to constuct list
def _list_append(mcoord, i_list, Zc):
for g in range(len(mcoord)):
grp = mcoord[g]
for i in range(len(grp)):
Z_p = grp[i]
if Z_p == Zc:
i_list.append((g, i))
# construct mutation list
for elem in self.element_count.iterkeys():
i_list = []
_Z = qtk.n2Z(elem)
_list_append(parent1['mutation'], i_list, _Z)
_list_append(parent2['mutation'], i_list, _Z)
constraint_list.append(i_list)
constraint_keys.append(elem)
while True:
selected_list = []
selected_coord = {}
consistant = True
for i in range(len(constraint_list)):
ind_location = constraint_list[i]
random.shuffle(ind_location)
elem = qtk.n2Z(constraint_keys[i])
_count = self.element_count[
constraint_keys[i]][0]
_end = _count
while len(set(ind_location[:_end])) < _count:
_end += 1
_new = list(set(ind_location[:_end]))
to_select = copy.deepcopy(selected_list)
to_select.extend(_new)
if len(set(to_select)) == \
len(selected_list) + len(_new):
selected_list.extend(_new)
for coord in _new:
selected_coord.update({coord: elem})
consistant = consistant & True
else:
consistant = False
child_ind = []
for g in range(len(parent1['mutation'])):
child_list = []
for i in range(len(parent1['mutation'][g])):
if selected_coord.has_key((g, i)):
child_list.append(selected_coord[(g,
i)])
else:
candidate = []
A1 = qtk.Z2n(parent1['mutation'][g][i])
A2 = qtk.Z2n(parent2['mutation'][g][i])
if not self.element_count.has_key(A1):
candidate.append(qtk.n2Z(A1))
if not self.element_count.has_key(A2):
candidate.append(qtk.n2Z(A2))
if len(candidate) == 0:
consistant = False
else:
child_list.append(
random.choice(candidate))
child_ind.append(child_list)
if consistant:
break
for g in range(len(parent1['mutation'])):
for i in range(len(parent1['mutation'][g])):
if not selected_coord.has_key((g, i)):
if random.random() < mutation_rate:
child_ind[g][i] = random.choice(\
self.mutation_target[g])
elem = qtk.Z2n(child_ind[g][i])
if self.element_count.has_key(elem):
coord = random.choice([tupl for tupl,targ
in selected_coord.iteritems()\
if targ == qtk.n2Z(elem)])
del selected_coord[coord]
selected_coord.update({
(g, i):
qtk.n2Z(elem)
})
[gc, ic] = list(coord)
new_targ = random.choice([z for z in \
self.mutation_target[gc]\
if not self.element_count\
.has_key(qtk.Z2n(z))])
child_ind[gc][ic] = new_targ
child.update({'mutation': child_ind})
else:
child_mut = []
for i in range(len(parent1[key])):
# mix two parent
grp = parent1[key][i]
index = range(len(grp))
random.shuffle(index)
ind1 = index[:len(index) / 2]
if len(index) % 2 == 0:
ind2 = index[len(index) / 2:]
append = -1
else:
ind2 = index[len(index) / 2:-1]
append = index[-1]
new = [parent1[key][i][ind] for ind in ind1]
new.extend([parent2[key][i][ind] for ind in ind2])
if append >= 0:
if random.random() >= 0.5:
new.append(parent1[key][i][append])
else:
new.append(parent2[key][i][append])
# mutation, loop through each element
for j in range(len(new)):
if random.random() < mutation_rate:
mut_target = [tar \
for tar in self.mutation_target[i]\
if tar != new[j]]
new[j] = random.choice(mut_target)
child_mut.append(new)
child.update({'mutation': child_mut})
else:
qtk.exit("ccs mate function is not yet implemented for "\
+ key)
return child
def findBonds(self, ratio=setting.bond_ratio, **kwargs):
del self.segments
del self.bond_types
self.segments = []
self.bond_types = {}
if 'no_report' not in kwargs or not kwargs['no_report']:
qtk.report("Molecule",
"finding bonds with cutoff ratio",
ratio)
def to_graph(l):
G = networkx.Graph()
for part in l:
# each sublist is a bunch of nodes
G.add_nodes_from(part)
# it also imlies a number of edges:
G.add_edges_from(to_edges(part))
return G
def to_edges(l):
"""
treat `l` as a Graph and returns it's edges
to_edges(['a','b','c','d']) -> [(a,b), (b,c),(c,d)]
"""
it = iter(l)
last = next(it)
for current in it:
yield last, current
last = current
itr = 0
bond_list = []
bonded = [False for i in range(self.N)]
for i in xrange(self.N):
for j in xrange(i+1, self.N):
d_ij = np.linalg.norm(self.R[i,:] - self.R[j,:])
atom_i = getattr(pt, self.type_list[i])
atom_j = getattr(pt, self.type_list[j])
Ri = atom_i.covalent_radius + \
atom_i.covalent_radius_uncertainty
Rj = atom_j.covalent_radius + \
atom_j.covalent_radius_uncertainty
Dij = (Ri+Rj) * float(ratio)
if d_ij < Dij:
bonded[i] = True
bonded[j] = True
if self.Z[i] < self.Z[j]:
atom_begin = self.Z[i]
atom_end = self.Z[j]
index_begin = i
index_end = j
else:
atom_begin = self.Z[j]
atom_end = self.Z[i]
index_begin = j
index_end = i
self.bonds[itr] = {'atom_begin' : atom_begin,
'index_begin' : index_begin,
'atom_end' : atom_end,
'index_end' : index_end,
'length' : d_ij}
bond_list.append([i, j])
type_begin = qtk.Z2n(atom_begin)
type_end = qtk.Z2n(atom_end)
bond_table = qtk.data.elements.bond_table
bond_keys = []
bond_keys = [
type_begin + _ + type_end for _ in ['-', '=', '#']
]
try:
bond_type_ind = np.argmin(
abs(
np.array([
bond_table[k][0] for k in bond_keys
if k in bond_table.keys()
]) - d_ij
)
)
except Exception as _e:
qtk.warning(
"error while processing bond" +\
str(bond_keys) + "with error message %s" % str(_e))
bond_type_ind = -1
bond_type = bond_keys[bond_type_ind]
self.bonds[itr]['name'] = bond_type
try:
bond_energy = \
bond_table[bond_keys[bond_type_ind]][1] * \
qtk.convE(1, 'kj-kcal')[0]
except:
bond_energy = np.nan
self.bonds[itr]['energy'] = bond_energy
if np.isnan(bond_energy):
qtk.warning("Non-tabliated covalent bond %s" % bond_type)
if bond_type in self.bond_types:
self.bond_types[bond_type] += 1
else:
self.bond_types[bond_type] = 1
itr += 1
segments = list(connected_components(to_graph(bond_list)))
for s in range(len(segments)):
segment = list(segments[s])
new_mol = self.getSegment(segment, **kwargs)
ns = len(self.segments)
new_mol.name = new_mol.name + '_%d' % ns
self.segments.append(new_mol)
for s in [i for i in range(self.N) if not bonded[i]]:
segment = [s]
new_mol = self.getSegment(segment, **kwargs)
ns = len(self.segments)
new_mol.name = new_mol.name + '_%d' % ns
self.segments.append(new_mol)
